MetCirc: navigating mass spectral similarity in high-resolution MS/MS metabolomics data

Mass spectrometry data is one of the key sources of information in many workflows in medicine and across the life sciences. Mass fragmentation spectra are considered characteristic signatures of the chemical compound they originate from, yet the chemical structure itself usually cannot be easily deduced from the spectrum. Often, spectral similarity measures are used as a proxy for structural similarity but this approach is strongly limited by a generally poor correlation between both metrics. Here, we propose MS2DeepScore: a novel Siamese neural network to predict the structural similarity between two chemical structures solely based on their MS/MS fragmentation spectra. Using a cleaned dataset of >100,000 mass spectra of about 15,000 unique known compounds, MS2DeepScore learns to predict structural similarity scores for spectrum pairs with high accuracy. In addition, sampling different model varieties through Monte-Carlo Dropout is used to further improve the predictions and assess the model's prediction uncertainty. On 3,600 spectra of 500 unseen compounds, MS2DeepScore is able to identify highly-reliable structural matches and predicts Tanimoto scores with a root mean squared error of about 0.15. The prediction uncertainty estimate can be used to select a subset of predictions with a root mean squared error of about 0.1. We demonstrate that MS2DeepScore outperforms classical spectral similarity measures in retrieving chemically related compound pairs from large mass spectral datasets, thereby illustrating its potential for spectral library matching. Finally, MS2DeepScore can also be used to create chemically meaningful mass spectral embeddings that could be used to cluster large numbers of spectra. Added to the recently introduced unsupervised Spec2Vec metric, we believe that machine learning-supported mass spectral similarity metrics have great potential for a range of metabolomics data processing pipelines.

Download Full-text

MS2DeepScore: a novel deep learning similarity measure to compare tandem mass spectra

Journal of Cheminformatics ◽

10.1186/s13321-021-00558-4 ◽

2021 ◽

Vol 13 (1) ◽

Author(s):

Florian Huber ◽

Sven van der Burg ◽

Justin J. J. van der Hooft ◽

Lars Ridder

Keyword(s):

Mass Spectra ◽

Mean Squared Error ◽

Similarity Measures ◽

Structural Similarity ◽

Metabolomics Data ◽

Spectral Similarity ◽

Mass Spectral ◽

Root Mean Squared Error ◽

Prediction Uncertainty ◽

Squared Error

AbstractMass spectrometry data is one of the key sources of information in many workflows in medicine and across the life sciences. Mass fragmentation spectra are generally considered to be characteristic signatures of the chemical compound they originate from, yet the chemical structure itself usually cannot be easily deduced from the spectrum. Often, spectral similarity measures are used as a proxy for structural similarity but this approach is strongly limited by a generally poor correlation between both metrics. Here, we propose MS2DeepScore: a novel Siamese neural network to predict the structural similarity between two chemical structures solely based on their MS/MS fragmentation spectra. Using a cleaned dataset of > 100,000 mass spectra of about 15,000 unique known compounds, we trained MS2DeepScore to predict structural similarity scores for spectrum pairs with high accuracy. In addition, sampling different model varieties through Monte-Carlo Dropout is used to further improve the predictions and assess the model’s prediction uncertainty. On 3600 spectra of 500 unseen compounds, MS2DeepScore is able to identify highly-reliable structural matches and to predict Tanimoto scores for pairs of molecules based on their fragment spectra with a root mean squared error of about 0.15. Furthermore, the prediction uncertainty estimate can be used to select a subset of predictions with a root mean squared error of about 0.1. Furthermore, we demonstrate that MS2DeepScore outperforms classical spectral similarity measures in retrieving chemically related compound pairs from large mass spectral datasets, thereby illustrating its potential for spectral library matching. Finally, MS2DeepScore can also be used to create chemically meaningful mass spectral embeddings that could be used to cluster large numbers of spectra. Added to the recently introduced unsupervised Spec2Vec metric, we believe that machine learning-supported mass spectral similarity measures have great potential for a range of metabolomics data processing pipelines.

Download Full-text

Searching a Comprehensive High Resolution Tandem Mass Spectral Library for Reliable Human Metabolite Identification

10.21748/am20.155 ◽

2020 ◽

Author(s):

Xiaoyu Yang ◽

Neta Pedatsur ◽

Stephen Stein

Keyword(s):

High Resolution ◽

Metabolite Identification ◽

Tandem Mass ◽

Mass Spectral Library ◽

Spectral Library ◽

Mass Spectral ◽

Human Metabolite

Download Full-text

A New Cucurbitane Glycoside from Siraitia grosvenorii

Natural Product Communications ◽

10.1177/1934578x1501000908 ◽

2015 ◽

Vol 10 (9) ◽

pp. 1934578X1501000 ◽

Cited By ~ 1

Author(s):

Venkata Sai Prakash Chaturvedula ◽

Srinivasa Rao Meneni

Keyword(s):

High Resolution ◽

Phytochemical Study ◽

Mass Spectral ◽

High Resolution Mass ◽

Siraitia Grosvenorii ◽

Commercial Extract ◽

Resolution Mass ◽

C Nmr

A systematic phytochemical study of the commercial extract of Luo Han Guo ( Siraitia grosvenorii) resulted in the isolation of an additional minor new cucurbitane glycoside, mogroside V A1 (1). The structure of the new compound was characterized on the basis of 1D (1H and 13C NMR) and 2D (COSY, HMQC, HMBC and NOESY) NMR and high resolution mass spectral (HRMS) data, as well as hydrolysis studies.

Download Full-text

Repository scale classification and decomposition of tandem mass spectral data

Scientific Reports ◽

10.1038/s41598-021-87796-6 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Mihir Mongia ◽

Hosein Mohimani

Keyword(s):

High Accuracy ◽

Mass Spectral Data ◽

Metabolomics Data ◽

Mass Spectral ◽

Molecular Features ◽

Oriented Samples ◽

Ingredient Composition ◽

Single Phenotype ◽

Tandem Mass Spectral Data ◽

Scale Classification

AbstractVarious studies have shown associations between molecular features and phenotypes of biological samples. These studies, however, focus on a single phenotype per study and are not applicable to repository scale metabolomics data. Here we report MetSummarizer, a method for predicting (i) the biological phenotypes of environmental and host-oriented samples, and (ii) the raw ingredient composition of complex mixtures. We show that the aggregation of various metabolomic datasets can improve the accuracy of predictions. Since these datasets have been collected using different standards at various laboratories, in order to get unbiased results it is crucial to detect and discard standard-specific features during the classification step. We further report high accuracy in prediction of the raw ingredient composition of complex foods from the Global Foodomics Project.

Download Full-text

The structure of novel C35 pentacyclic terpenes from Acetobacter xylinum

Biochemical Journal ◽

10.1042/bj1350133 ◽

1973 ◽

Vol 135 (1) ◽

pp. 133-143 ◽

Cited By ~ 63

Author(s):

Hans J. Förster ◽

Klaus Biemann ◽

W. Geoffrey Haigh ◽

Neil H. Tattrie ◽

J. Ross Colvin

Keyword(s):

High Resolution ◽

High Resolution Mass Spectrometry ◽

Carbon Chain ◽

Acetobacter Xylinum ◽

Mass Spectrometric ◽

Hydroxyl Groups ◽

Mass Spectral Data ◽

Mass Spectral ◽

High Resolution Mass ◽

Resolution Mass

A novel C35 terpene and its monounsaturated analogue were isolated from cultures of Acetobacter xylinum, together with traces of their C36 homologues. These substances were found to be hopane derivatives substituted by a five-carbon chain bearing four vicinal hydroxyl groups. For the parent hydrocarbon the term bacteriohopane is proposed. The elucidation of the structures utilized high-resolution mass spectrometry of the terpenes, degradation to C32 hydrocarbons and detailed mass-spectrometric comparison of these with C32 hydrocarbons synthesized from known pentacyclic triterpenes. High-resolution mass-spectral data of the terpenes are presented. N.m.r. data are in agreement with the proposed structures, which are further supported by the isolation from the same organism of 22-hydroxyhopane and derivative hopene(s).

Download Full-text

Mass spectral studies of perfluorooctane sulfonate derivatives separated by high-resolution gas chromatography

Rapid Communications in Mass Spectrometry ◽

10.1002/rcm.3241 ◽

2007 ◽

Vol 21 (22) ◽

pp. 3547-3553 ◽

Cited By ~ 31

Author(s):

Ingrid Langlois ◽

Urs Berger ◽

Zdenek Zencak ◽

Michael Oehme

Keyword(s):

Gas Chromatography ◽

High Resolution ◽

Perfluorooctane Sulfonate ◽

Spectral Studies ◽

Mass Spectral ◽

High Resolution Gas Chromatography

Download Full-text

A scalable workflow for the human exposome

10.21203/rs.3.rs-83843/v1 ◽

2020 ◽

Author(s):

Xin Hu ◽

Douglas Walker ◽

YongLiang Liang ◽

Matthew Smith ◽

Michael Orr ◽

...

Keyword(s):

High Resolution ◽

High Resolution Mass Spectrometry ◽

A Priori ◽

Population Level ◽

Sample Volume ◽

Single Step ◽

Environmental Chemicals ◽

Environmental Chemical ◽

Mass Spectral ◽

Sensitivity And Selectivity

Abstract Complementing the genome with an understanding of the human exposome is an important challenge for contemporary science and technology. Tens of thousands of chemicals are used in commerce, yet cost for targeted environmental chemical analysis limits surveillance to a few hundred known hazards. To overcome limitations which prevent scaling to thousands of chemicals, we developed a single-step express liquid extraction (XLE), gas chromatography high-resolution mass spectrometry (GC-HRMS) analysis and computational pipeline to operationalize the human exposome. We show that the workflow supports quantification of environmental chemicals in small human plasma (200 µL) and tissue (≤ 100 mg) samples. The method also provides high resolution, sensitivity and selectivity for exposome epidemiology of mass spectral features without a priori knowledge of chemical identity. The simplicity of the method can facilitate harmonization of environmental biomonitoring between laboratories and enable population level human exposome research with limited sample volume.

Download Full-text